15  FEB

February 15, 2022

Applications of AI in Cell Segmentation

Cell segmentation is a method for quantifying the number of cells within a tissue section or a given area. Cell segmentation is the basis of all tissue cytometry analyses, including FISH/CISH/RNAscope detection, in-depth phenotyping, and spatial analysis. Therefore, accurate cell segmentation is crucial for most tissue cytometry approaches in research and clinics.

There are a number of different experimental approaches to performing cell segmentation. Despite its importance in many medical and research applications, cell segmentation is not necessarily a straightforward task, and issues and deficiencies with certain methodologies have led to long-term misunderstandings about the ratios of particular cell types in organs. ¹

Some challenges in retrieving accurate cell counts include difficulties discriminating between cells and debris, appropriate samples, standardized staining protocols, and human errors. There are also limited validation processes for cell segmentation.

As well as increasing throughput for cell segmentation, all of the inaccuracies in current classical approaches may be addressed by moving towards AI-based approaches. 

Image Recognition and Deep Learning

One of the most common AI approaches for cell counting is to use deep learning with convolutional neural networks (CNNS). ² Deep learning methods are becoming increasingly popular as there are more ways to automate model training, which reduces another laboratory-intensive step in the preparation. ³

Deep learning-based cell segmentation approaches work with minimal user input for a variety of different cell types and can also be interfaced with more complex multispectral cytometry approaches. Using multiple tags in microscopy imaging can help improve confidence in assigning certain cellular phenotypes, but the analysis and understanding of such images and the complex datasets they produce can be challenging.

Tissue Gnostics and Cell Segmentation

With AI processes the entire cell segmentation procedure can be improved, even for these more complex cytometry methods. TissueGnostics has long-standing expertise in developing automated software solutions for image analysis, with cell segmentation being no exception. This can be combined with all other kinds of analysis, such as tissue classification and spatial analysis or dot detection (FISH, CISH, RNAScope).

Contact TissueGnostics today to find out how analysis workflows, including deep-learning-based cell segmentation, could improve your throughput and the accuracy of your cell image analysis. With AI algorithms becoming easier to use and more computationally efficient, introducing standardized analysis procedures for cell segmentation has never been easier and could enhance the reliability of your data.

References and Further Reading

1. by Bartheld, CS, Bahney, J., & Herculano-Houzel, S. (2016). The search for true numbers of neurons and glial cells in the human brain: A review of 150 years of cell counting. Journal of Comparative Neurology, 524(18), 3865–3895. https://doi.org/10.1002/cne.24040
2. Pan, X., Yang, D., Li, L., Liu, Z., Yang, H., Cao, Z., He, Y., Ma, Z., & Chen, Y. (2018). Cell detection in pathology and microscopy images with multi-scale fully convolutional neural networks. World Wide Web, 21(6), 1721-1743. https://doi.org/10.1007/s11280-017-0520-7
3. Zhan, G., Wang, W., Sun, H., Hou, Y., & Feng, L. (2022). Auto-CSC: A Transfer Learning Based Automatic Cell Segmentation and Count Framework. Cyborg and Bionic Systems, 2022, 1–10. https://doi.org/10.34133/2022/9842349